Difference in differences for stayers with a time-varying qualification: Health expenditure elasticity of the elderly

Myoung-jae Lee, Young Sook Kim

Research output: Contribution to journalArticle

Abstract

In difference in differences, a treatment is applied only to a qualified group at some time point. The qualification may be time-constant as in gender, or time-varying as in residential location. When the qualification is time-varying, there appear four groups: the newly qualified (in-movers), the already qualified (in-stayers), the newly disqualified (out-movers), and the already disqualified (out-stayers). A change in qualification may affect the response variable of interest even when the treatment effect is zero, which is an 'untreated moving effect'. Also, when the treatment effect is not zero, it may be different across the four groups. The conventional difference in differences fails to remove untreated moving effects and ignores the possible treatment effect heterogeneity across the groups. This paper shows how to account for untreated moving effects and proposes 'the effect on in-stayers' as the main effect of interest. Our proposal can be implemented with least squares estimator for panel models or with nonparametric methods. An empirical analysis is provided using Korean data for the effects of the basic elder pension on health-care expenditure.

Original languageEnglish
Pages (from-to)1134-1145
Number of pages12
JournalHealth Economics (United Kingdom)
Volume23
Issue number9
DOIs
Publication statusPublished - 2014 Jan 1

Fingerprint

Elasticity
Health Expenditures
Pensions
Therapeutics
Least-Squares Analysis
Delivery of Health Care

ASJC Scopus subject areas

  • Health Policy

Cite this

Difference in differences for stayers with a time-varying qualification : Health expenditure elasticity of the elderly. / Lee, Myoung-jae; Kim, Young Sook.

In: Health Economics (United Kingdom), Vol. 23, No. 9, 01.01.2014, p. 1134-1145.

Research output: Contribution to journalArticle

@article{9a1e83c275a6433a9096c0f25e26f723,
title = "Difference in differences for stayers with a time-varying qualification: Health expenditure elasticity of the elderly",
abstract = "In difference in differences, a treatment is applied only to a qualified group at some time point. The qualification may be time-constant as in gender, or time-varying as in residential location. When the qualification is time-varying, there appear four groups: the newly qualified (in-movers), the already qualified (in-stayers), the newly disqualified (out-movers), and the already disqualified (out-stayers). A change in qualification may affect the response variable of interest even when the treatment effect is zero, which is an 'untreated moving effect'. Also, when the treatment effect is not zero, it may be different across the four groups. The conventional difference in differences fails to remove untreated moving effects and ignores the possible treatment effect heterogeneity across the groups. This paper shows how to account for untreated moving effects and proposes 'the effect on in-stayers' as the main effect of interest. Our proposal can be implemented with least squares estimator for panel models or with nonparametric methods. An empirical analysis is provided using Korean data for the effects of the basic elder pension on health-care expenditure.",
keywords = "difference in differences, effect on in-movers, effect on in-stayers, health expenditure, panel data, untreated moving effect",
author = "Myoung-jae Lee and Kim, {Young Sook}",
year = "2014",
month = "1",
day = "1",
doi = "10.1002/hec.3049",
language = "English",
volume = "23",
pages = "1134--1145",
journal = "Health Economics (United Kingdom)",
issn = "1057-9230",
publisher = "John Wiley and Sons Ltd",
number = "9",

}

TY - JOUR

T1 - Difference in differences for stayers with a time-varying qualification

T2 - Health expenditure elasticity of the elderly

AU - Lee, Myoung-jae

AU - Kim, Young Sook

PY - 2014/1/1

Y1 - 2014/1/1

N2 - In difference in differences, a treatment is applied only to a qualified group at some time point. The qualification may be time-constant as in gender, or time-varying as in residential location. When the qualification is time-varying, there appear four groups: the newly qualified (in-movers), the already qualified (in-stayers), the newly disqualified (out-movers), and the already disqualified (out-stayers). A change in qualification may affect the response variable of interest even when the treatment effect is zero, which is an 'untreated moving effect'. Also, when the treatment effect is not zero, it may be different across the four groups. The conventional difference in differences fails to remove untreated moving effects and ignores the possible treatment effect heterogeneity across the groups. This paper shows how to account for untreated moving effects and proposes 'the effect on in-stayers' as the main effect of interest. Our proposal can be implemented with least squares estimator for panel models or with nonparametric methods. An empirical analysis is provided using Korean data for the effects of the basic elder pension on health-care expenditure.

AB - In difference in differences, a treatment is applied only to a qualified group at some time point. The qualification may be time-constant as in gender, or time-varying as in residential location. When the qualification is time-varying, there appear four groups: the newly qualified (in-movers), the already qualified (in-stayers), the newly disqualified (out-movers), and the already disqualified (out-stayers). A change in qualification may affect the response variable of interest even when the treatment effect is zero, which is an 'untreated moving effect'. Also, when the treatment effect is not zero, it may be different across the four groups. The conventional difference in differences fails to remove untreated moving effects and ignores the possible treatment effect heterogeneity across the groups. This paper shows how to account for untreated moving effects and proposes 'the effect on in-stayers' as the main effect of interest. Our proposal can be implemented with least squares estimator for panel models or with nonparametric methods. An empirical analysis is provided using Korean data for the effects of the basic elder pension on health-care expenditure.

KW - difference in differences

KW - effect on in-movers

KW - effect on in-stayers

KW - health expenditure

KW - panel data

KW - untreated moving effect

UR - http://www.scopus.com/inward/record.url?scp=84906355551&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906355551&partnerID=8YFLogxK

U2 - 10.1002/hec.3049

DO - 10.1002/hec.3049

M3 - Article

C2 - 24733617

AN - SCOPUS:84906355551

VL - 23

SP - 1134

EP - 1145

JO - Health Economics (United Kingdom)

JF - Health Economics (United Kingdom)

SN - 1057-9230

IS - 9

ER -